feat(mcp): expose server prompts as dynamic skills - #1813
Open
Aaronontheweb wants to merge 8 commits into
Open
Conversation
| // Resolve the full path and verify it's within the skill directory | ||
| var fullPath = Path.GetFullPath(Path.Combine(skill.SkillDirectory, resourcePath)); | ||
| var skillDirFull = Path.GetFullPath(skill.SkillDirectory); | ||
| var fullPath = Path.GetFullPath(Path.Combine(fileSource.SkillDirectory, resourcePath)); |
Comment on lines
+74
to
+79
| [Path.Combine( | ||
| SmokeMcpServerLocator.LocateRepositoryRoot(), | ||
| "evals", | ||
| "fixtures", | ||
| "mcp", | ||
| "prompt_server.py")], |
| Assert.NotNull(repo); | ||
|
|
||
| var projectDir = Path.Combine(repo!.FullName, "tests", "Netclaw.SmokeMcpServer"); | ||
| var projectDir = Path.Combine(LocateRepositoryRoot(), "tests", "Netclaw.SmokeMcpServer"); |
| public static string LocateRepositoryRoot() | ||
| { | ||
| var repo = new DirectoryInfo(AppContext.BaseDirectory); | ||
| while (repo is not null && !File.Exists(Path.Combine(repo.FullName, "Netclaw.slnx"))) |
Collaborator
Author
Adversarial reviewI reviewed the protocol, permission, lifecycle, service construction, output, and compatibility paths. Fixed findings
Deferred hardening
Verified boundaries
Verification after review fixes
The model behavior eval remains incomplete. The environment lacks the required evaluator provider variables. |
Collaborator
Author
|
CI found a stale native smoke assertion. The daemon now reports the full MCP catalog as Commit |
Aaronontheweb
force-pushed
the
feature/mcp-prompt-skills
branch
from
August 9, 2026 01:36
2e08343 to
9671f97
Compare
Comment on lines
+232
to
+241
| foreach (var skill in promptSkills) | ||
| { | ||
| if (skill.Source is not McpPromptSkillSource source | ||
| || !string.Equals(source.ServerName, serverName, StringComparison.OrdinalIgnoreCase)) | ||
| { | ||
| throw new ArgumentException( | ||
| $"Skill '{skill.Name}' is not an MCP prompt from server '{serverName}'.", | ||
| nameof(skills)); | ||
| } | ||
| } |
Aaronontheweb
force-pushed
the
feature/mcp-prompt-skills
branch
from
August 9, 2026 13:20
9671f97 to
cbca71a
Compare
Reject cross-server logical-name conflicts before catalog publication. Fix generated string-map binding and add direct permission tests.
Aaronontheweb
force-pushed
the
feature/mcp-prompt-skills
branch
from
August 9, 2026 14:20
cbca71a to
de1b124
Compare
Aaronontheweb
marked this pull request as ready for review
August 9, 2026 14:36
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
skill_loadwith validated arguments.Scope
This PR implements issue #1806 from epic #1805.
MCP resource reads remain in #1807. Proactive catalog subscriptions remain in #1808.
Verification
dotnet build Netclaw.slnx --no-restore --nologo -v:minimaldotnet slopwatch analyzepwsh ./scripts/Add-FileHeaders.ps1 -Verifyopenspec validate add-mcp-prompt-skills --strictgit diff --checkThe model behavior eval did not run. The environment lacks the required evaluator provider variables.
Closes #1806.